Automatic Inspection of TFT-LCD Glass Substrates Using Optimized Support Vector Machines

Yousefian-Jazi, Ali; Liu, J. Jay; Lee, Hokyung

doi:10.3182/20120710-4-sg-2026.00054

Cited by 14 publications

(13 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Once this is done, the different metaheuristics are implemented to optimize the training of a one-class classifier (SVDD), that is, the hyper-parameters are optimally found that improve the abilities of SVDD for faults detection in industrial processes where there is not much faulty information available. In particular, advantages of the exploration and exploitation of metaheuristic algorithms are taken in order to tune the hyper-parameters C in SVDD and s for the RBF kernel (16). Once the SVDD training is optimized, it will be ready to monitor new data and detect possible faults in the multivariate industrial process.…”

Section: Methodology For Fault Detectionmentioning

confidence: 99%

See 1 more Smart Citation

Metaheuristics and Support Vector Data Description for Fault Detection in Industrial Processes

et al. 2020

View full text Add to dashboard Cite

In this study, a system for faults detection using a combination of Support Vector Data Description (SVDD) with metaheuristic algorithms is presented. The presented approach is applied to a real industrial process where the set of measured faults is scarce. The original contribution in this work is the industrial context of application and the comparison of swarm intelligence algorithms to optimize the SVDD hyper-parameters. Four recent metaheuristics are compared hereby to solve the corresponding optimization problem in an efficient manner. These optimization techniques are then implemented for fault detection in a multivariate industrial process with non-balanced data. The obtained numerical results seem to be promising when the considered optimization techniques are combined with SVDD. In particular, the Spotted Hyena algorithm outperforms other metaheuristics reaching values of F1 score near 100% in fault detection.

show abstract

Section: Methodology For Fault Detectionmentioning

confidence: 99%

“…There are some studies covering different SVDD algorithm applications [15][16][17]-most of them optimizing the hyper-parameters using approaches like grid search, which is computationally expensive. To obtain these parameters in a more efficient manner, some authors have considered metaheuristic algorithms.…”

Section: Introductionmentioning

confidence: 99%

Metaheuristics and Support Vector Data Description for Fault Detection in Industrial Processes

et al. 2020

View full text Add to dashboard Cite

show abstract

“…Jazi et al [2] proposed an automatic inspection method with support vector machine (SVM) and simulated anneal (SA) to identify the Mura defects for the TFT-LCD products. This method used SA and parallel genetic algorithm to reduce the number of features gaining 15.5% more accuracy than the result of SVM without feature selection.…”

Section: A Repair Improvement In Manufacturingmentioning

confidence: 99%

“…In thin-film transistor liquid-crystal display (TFT-LCD) industry, computer vision techniques have been widely used to help manufacturers monitor abnormalities, identify potential process bottlenecks, and swiftly respond to process problems to reduce yield loss. In the earlier years, automatic optical inspection (AOI) machines have achieved satisfactory functional defect identification [1], [2]. However, it is difficult to identify irregular defect such as Mura by using typical AOI [3].…”

Section: Introductionmentioning

confidence: 99%

A Novel Attention-Based Multi-Modal Modeling Technique on Mixed Type Data for Improving TFT-LCD Repair Process

Lai

2022

IEEE Access

View full text Add to dashboard Cite

In the Thin-Film Transistor Liquid-Crystal Display (TFT-LCD) manufacturing, conducting a machine learning based system with multiple data types has become urgently desired to solve complicated problems. This paper proposes a novel deep learning approach: TabVisionNet, which is modeled by utilizing the information from both tabular data and image data. Tabular data and image data are first encoded by a Tabular encoder and a CNN encoder respectively, then a novel attention mechanism called Sequential Decision Attention was integrated into the framework, that combines the information from two modalities. This novel approach can capture the complex relationship between modalities then gain better generalization and faster convergence in the training process. Conducting an experiment, the performance of our novel approach was significantly better than single-modal and other multi-modal learning approaches in our real case scenario.

show abstract

“…FPDs are usually large, but the size of defects on them is of the order of hundreds of micrometers. As such, the manual inspection of FPDs is not suitable [3].…”

Section: Introductionmentioning

confidence: 99%

Optimal Color Lighting for Scanning Images of Flat Panel Display using Simplex Search

Kim

Jin

et al. 2018

J. Imaging

View full text Add to dashboard Cite

A system for inspecting flat panel displays (FPDs) acquires scanning images using multiline charge-coupled device (CCD) cameras and industrial machine vision. Optical filters are currently installed in front of these inspection systems to obtain high-quality images. However, the combination of optical filters required is determined manually and by using empirical methods; this is referred to as passive color control. In this study, active color control is proposed for inspecting FPDs. This inspection scheme requires the scanning of images, which is achieved using a mixed color light source and a mixing algorithm. The light source utilizes high-power light emitting diodes (LEDs) of multiple colors and a communication port to dim their level. Mixed light illuminates an active-matrix organic light-emitting diode (AMOLED) panel after passing through a beam expander and after being shaped into a line beam. The image quality is then evaluated using the Tenenbaum gradient after intensity calibration of the scanning images. The dimming levels are determined using the simplex search method which maximizes the image quality. The color of the light was varied after every scan of an AMOLED panel, and the variation was iterated until the image quality approached a local maximization. The number of scans performed was less than 225, while the number of dimming level combinations was 2048 4 . The proposed method can reduce manual tasks in setting-up inspection machines, and hence is useful for the inspection machines in FPD processes.

show abstract

Automatic Inspection of TFT-LCD Glass Substrates Using Optimized Support Vector Machines

Cited by 14 publications

References 26 publications

Metaheuristics and Support Vector Data Description for Fault Detection in Industrial Processes

Metaheuristics and Support Vector Data Description for Fault Detection in Industrial Processes

A Novel Attention-Based Multi-Modal Modeling Technique on Mixed Type Data for Improving TFT-LCD Repair Process

Optimal Color Lighting for Scanning Images of Flat Panel Display using Simplex Search

Contact Info

Product

Resources

About